If they had done a pilot program with one-tenth the number of cars, they still would have had an average of 40 cars per state. Wouldn't that have been enough to qualify customer interest and demand? Did they really need to gamble a third of a trillion dollars just to test the market?
You're forgiven because the article and summary are clickbait garbage, but these Teslas were primarily for Uber drivers, and the demand is there. The problem is with high repair costs, and rideshare drivers specifically. They downsized that side despite the demand, to limit costs, and the other side of business that you're thinking of wound up with too many spilled over. Some were supposed to go back after insurance changes for the rideshare side, and a bunch were to be sold off.